Detection of Low Level Nilotinib or Dasatinib Resistant BCR-ABL Mutations by Mass Spectrometry In CML Patients Who Fail Imatinib Is Highly Predictive of Their Subsequent Clonal Expansion When Treated with the Drug for Which Their Mutation Confers Resistance

Abstract 891

Imatinib (IM) resistance is commonly associated with the acquisition of BCR-ABL kinase domain (KD) mutations. Nilotinib (NIL) and dasatinib (DAS) are active against the majority of IM resistant mutations, however a small number confer clinical resistance to NIL (Y253H, E255K/V and F359V/C) or DAS (V299L, T315A and F317L/I), or both (T315I). These mutations are associated with low response rates and their detection after IM failure aids selection of the most appropriate therapy; however, mutations may be present below the detection limit of conventional direct sequencing (dir-seq, sensitivity 10–20%). We aimed to determine whether more sensitive detection of NIL/DAS resistant mutations prior to commencing NIL or DAS therapy in patients (pts) who failed IM could predict their subsequent clonal expansion in the presence of the inhibitor for which they confer resistance, and whether this affects response. These mutations will be referred to as ‘inhibitor resistant’ when detected in a patient subsequently treated with the inhibitor for which the mutation confers resistance, and conversely, ‘inhibitor sensitive’ when detected in a pt treated with the inhibitor which retains activity against the mutation. We developed 4 multiplexed genotyping assays using high throughput chip-based mass spectrometry (Sequenom MassARRAY; M-A) to detect 27 mutations that account for approx. 88% of all mutations, including all of the NIL/DAS resistant mutations (sensitivity 0.05–0.5%). Samples of 210 CML pts treated with NIL (n=85) or DAS (n=125) after IM failure (CP n=102, AP n=64, BC n=44) were retrospectively analyzed by M-A at baseline (before commencing NIL or DAS), and by dir-seq at baseline and during follow up (FU) (median FU 12 mo, r 1–36). When a mutation was detected by dir-seq, the concordance of detection by M-A was 99.4%. Dir-seq detected 26 inhibitor resistant mutations at baseline in 24 pts. However, M-A detected 22 additional inhibitor resistant mutations in 19 pts (CP 7/102 7%, AP 8/64 13%, MBC 4/44 9%). Twelve of the 22 mutations were T315I and 11 (92%) subsequently became a dominant clone detectable by dir-seq (median 3 mo, r 1–12). Of the remaining 10 inhibitor resistant mutations, 7 (70%) became dominant (median 7.5 mo, r 1–24). The exceptions were F317L in 2 pts and E255K in 1. In these 3 pts, T315I was also present at baseline (1) or became dominant at 3 or 9 mo (2), suggesting a hierarchy of emergent resistant mutations according to mutant drug sensitivity. A complete cytogenetic response (CCyR, or its BCR-ABL equivalent of <1% IS) was achieved by 0/25 pts with inhibitor resistant mutations detectable at baseline by dir-seq, and by 2/19 (10%; CP n=1, AP n=1) pts with inhibitor resistant mutations detected only by M-A (3 pts overlapped both groups). In contrast, 170 pts had no inhibitor resistant mutations detected at baseline by dir-seq or M-A and 64 (38%) achieved CCyR (P<.001; 87/170 pts were CP and 44% achieved CCyR). In addition to the mutations already discussed, 22 NIL/DAS resistant mutations were detected only by M-A in 17/210 pts at baseline who were not subsequently treated with the inhibitor for which the mutation confers resistance (inhibitor sensitive mutations), e.g. F317L in a pt subsequently treated with NIL. During FU, 3/22 inhibitor sensitive mutations subsequently emerged in 2 pts, but 19 never emerged (86%; median FU 9 mo, r 1–30). Importantly, none of the 27 mutations in the M-A assay were detected in the ABL KD of 30 normal donors or 20 replicates of a BCR-ABL-negative cell line. In conclusion, our multiplex M-A assay, with sensitivity to 0.05%, could simultaneously detect the bulk of IM resistant mutations and predict the subsequent clonal expansion of resistant mutations during NIL/DAS therapy with high confidence. Overall, 82% of the inhibitor resistant mutations detected only by M-A at baseline became a dominant mutation. In contrast, only 14% of the inhibitor sensitive mutations detected only by M-A at baseline became detectable by dir-seq, P<.0001 (table). Sensitive detection of NIL/DAS resistant mutations after IM failure offers vital information when considering therapeutic options.